On Tue, 2009-07-14 at 12:21 +0100, Spencer Oliver wrote:
> >
> > In this respect, I think all of the drivers should be enabled 
> > by default.  Developers could disable them by choice, but 
> > their underlying libraries are being detected such that we 
> > can add new checks to disable the drivers when their 
> > dependencies are missing.
> > 
> > The goal is to allow our users to simply run './configure' 
> > without having to give any messy options.  Everything should 
> > be detected automatically, picking "sane" default settings if 
> > possible.  While that may not be 100% possible, we should be 
> > able to do _much_ better than the 0.2.0 configure script 
> > manages to accomplish.
> >
> 
> That sounds like a good plan, urjtag use a similar behaviour for detecting
> usb stuff.
>  
> > > Surely let the developer enable what he wants?
> > 
> > The non-x86 developer does not have a choice; the configure 
> > script always forced it to be enabled.  The developer still 
> > has the choice to disable it, so a choice is still there -- 
> > it's just backwards now.
> > It's broken either way, really.
> > 
> > Clearly, the configure script needs to be revisited during 0.3.0.
> > We need to figure out when and where these options are 
> > required, and we can test in the script for those required 
> > conditions.  These is are easy to script, once we know what they are.
> > 
> 
> attached patch fixes ppdev and win32 platforms - no objections and i will
> commit?

A "real" fix will require more work, but this fix could be merged for
0.2.1 -- if we need to produce such a release.  This issue is not
sufficient to demand such given there is an easy workaround, but I am
very sorry that I caused it at the last minute before the release.
To be safe, I should have changed the documentation to reflect the code.

--Z
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to